This page explains our data and consent practices in plain language. It is not a substitute for the full Privacy Policy or the specific consent form each contributor signs, and it should be reviewed by legal counsel before being treated as a binding policy.
Who provides the data
Recordings in our archives, including ALPIA, come from individuals who choose to take part as paid contributors. We do not scrape audio or text from the internet, social media or call recordings without explicit, individual consent.
What contributors agree to
Before recording anything, each contributor is told, in a language they understand:
- What will be recorded and in what context (words, phrases, conversation or a specific domain such as healthcare or customer service).
- Who will be able to access the recording: our internal team, a specific client, the wider research community or the public, depending on the project.
- Whether the recording may be used to train commercial models, and by whom.
- How long we intend to retain the recording and how to request its removal.
- That taking part is voluntary and that payment is not conditional on continuing beyond any point they choose to stop.
Payment
Contributors are paid for their time and their recordings. Payment terms are agreed before recording begins and are not reduced or withheld based on how a recording is later used, within the terms the contributor originally agreed to.
What we collect about a speaker
We collect only the metadata that makes a recording scientifically or technically useful: language, variety or dialect, region, broad age band, and general recording conditions. We do not require full names, identification numbers or contact details to be attached to a public dataset entry.
Dialect and variety
We record how people actually speak, including regional and national varieties, code-switching and informal speech. This is treated as valuable data, not something to be corrected or standardised away.
Client, research and internal use
Some datasets are collected specifically for a paying client and are not shared beyond that engagement unless the contributor agreed otherwise. Other datasets are collected for research or archive purposes and may be more widely available under the access tiers described on the ALPIA page. Where a dataset also supports our own internal research and development, this is disclosed to contributors as part of the consent process, not treated as a separate hidden use.
Your rights as a contributor
If you have contributed a recording and want to know how it is being used, request its removal, or ask a question about consent, contact us using the details below.
Questions or requests
Email languages@grhoglobal.com with the subject line "Data and consent" and we will respond within a reasonable time.
